Arabinose biosynthesis is critical for salt stress tolerance inArabidopsis

Chunzhao Zhao1,2* , Omar Zayed2* , Fansuo Zeng3, Chaoxian Liu4, Ling Zhang5 , Peipei Zhu6 ,

Chuan-Chih Hsu6 , Yunus E. Tuncil7, W. Andy Tao6, Nicholas C. Carpita8,9 and Jian-Kang Zhu1,2

1CAS Center for Excellence in Molecular Plant Sciences, Shanghai Center for Plant Stress Biology, Chinese Academy of Sciences, Shanghai 201602, China; 2Department of Horticulture and

Landscape Architecture, Purdue University, West Lafayette, IN 47907, USA; 3State Key Laboratory of Tree Genetics and Breeding, Northeast Forestry University, Harbin 150040, China;

4Maize Research Institute, Southwest University, Chongqing 400715, China; 5Jilin Provincial Key laboratory of Agricultural Biotechnology, Jilin Academy of Agricultural Sciences, Changchun,

Jilin 130033, China; 6Department of Biochemistry, Purdue University, West Lafayette, IN 47907, USA; 7Food Engineering Department, Ordu University, Ordu 52200, Turkey; 8Department

of Botany and Plant Pathology, Purdue University, West Lafayette, IN 47907, USA; 9Purdue Center for Plant Biology, Purdue University, West Lafayette, IN 47907, USA

Authors for correspondence:Chunzhao Zhao

Tel: +86 21 57078274Email: czzhao@psc.ac.cn

Jian-Kang Zhu

Tel: +86 21 57078201Email: jkzhu@sibs.ac.cn

Received: 30 December 2018

Accepted: 16 April 2019

New Phytologist (2019)doi: 10.1111/nph.15867

Key words: Arabidopsis, arabinogalactanprotein, arabinose, cell wall integrity, rootelongation, salt stress.

Summary

� The capability to maintain cell wall integrity is critical for plants to adapt to unfavourable

conditions. L-Arabinose (Ara) is a constituent of several cell wall polysaccharides and many cell

wall-localised glycoproteins, but so far the contribution of Ara metabolism to abiotic stress

tolerance is still poorly understood.� Here, we report that mutations in the MUR4 (also known as HSR8) gene, which is required

for the biosynthesis of UDP-Arap in Arabidopsis, led to reduced root elongation under high

concentrations of NaCl, KCl, NaNO3, or KNO3.� The short root phenotype of the mur4/hsr8 mutants under high salinity is rescued by

exogenous Ara or gum arabic, a commercial product of arabinogalactan proteins (AGPs) from

Acacia senegal. Mutation of the MUR4 gene led to abnormal cell�cell adhesion under salt

stress. MUR4 forms either a homodimer or heterodimers with its isoforms. Analysis of the

higher order mutants of MUR4 with its three paralogues, MURL, DUR, MEE25, reveals that

the paralogues of MUR4 also contribute to the biosynthesis of UDP-Ara and are critical for

root elongation.� Taken together, our work revealed the importance of the Ara metabolism in salt stress toler-

ance and also provides new insights into the enzymes involved in the UDP-Ara biosynthesis in

plants.

Introduction

L-Arabinose (Ara) is a plant-specific monosaccharide that occursin 10–20% of noncellulosic cell wall polysaccharides in Ara-bidopsis (Rautengarten et al., 2017). Ara is found in arabinansand type I arabinogalactans, which are the major components ofthe pectic wall polymer rhamnogalacturonan-I (RG-I), and con-stitutes critical residues in the rhamnogalacturonan II (RG-II)variant of homogalacturonan for cross-linking with boron (Caf-fall & Mohnen, 2009). Ara residues are found in hemicellulosicglucuronoarabinoxylan (GAX) and the xyloglucan of Solanaceousspecies (Carpita & Gibeaut, 1993; Pe~na et al., 2008; Scheller &Ulvskov, 2010; Schultink et al., 2013, 2014). In addition topolysaccharides, many glycoproteins, such as extensins andleucine-rich repeat extensins, some secreted CLE peptides, andubiquitous proteoglycan arabinogalactan proteins (AGPs), arealso arabinosylated (Showalter, 1993; Velasquez et al., 2011;Okamoto et al., 2013; Borassi et al., 2016; Marzol et al., 2018).

Ara has two ring forms, the furanose (Araf) and the pyranose(Arap) form. In most Ara-containing components in plants, Araoccurs in the furanose form. However, Ara is initially synthesizedas a form of UDP-Arap either via a de novo pathway or via a sal-vage pathway, then UDP-Arap is converted to UDP-Araf in thecytosol (Konishi et al., 2007; Kotake et al., 2016). The de novosynthesis of UDP-Arap is initiated from the conversion of UDP-glucose (UDP-Glc) to UDP-glucuronic acid (UDP-GlcA) catal-ysed by UDP-Glc dehydrogenases (UGDs; Klinghammer &Tenhaken, 2007); UDP-glucuronic acid decarboxylases (UXSs)subsequently catalyse the conversion of UDP-GlcA to UDP-xylose (UDP-Xyl; Kuang et al., 2016). The conversion of UDP-Xyl to UDP-Arap is catalysed by UDP-Xyl 4-epimerases (UXEs)in the Golgi lumen (Burget et al., 2003). The synthesized UDP-Arap is transported from the Golgi lumen to the cytosol, where aUDP-Ara mutase (UAM) interconverts UDP-Arap to UDP-Araf,which is transported back into the Golgi lumen to be incorpo-rated into Ara-containing polysaccharides (Konishi et al., 2007;Rautengarten et al., 2017). In Arabidopsis, the mutant murus4

*These authors contributed equally to this work.

� 2019 The Authors

New Phytologist� 2019 New Phytologist Trust

New Phytologist (2019) 1www.newphytologist.com

Research

(mur4) with a lower level of cell wall Ara was shown be defectivein UXE1 (Burget & Reiter, 1999; Burget et al., 2003). In themur4 mutant, the Ara content in the cell wall of rosette leaves isreduced by c. 50% (Burget & Reiter, 1999; Li et al., 2007). That50% of the Ara remains in the mur4 mutant indicated that otherenzymes were involved in the biosynthesis of UDP-Ara.

A salvage pathway recycles Ara that is released during theturnover of Ara-containing polymers and glycoproteins in the cellwall (Dolezal & Cobbett, 1991; Sherson et al., 1999; Behm€ulleret al., 2016). Free arabinoses are phosphorylated by two C-1 arabi-nokinases, ARA1 and ARA2, and are then converted to UDP-Arap by a UDP-sugar pyrophosphorylase (NDP) with a broadsubstrate specificity (Kotake et al., 2004, 2007; Behm€uller et al.,2016). Mutations in ARA1 or in both the ARA1 and ARA2 genesresult in an increase in cytoplasmic Ara content compared withthe wild-type (Dolezal & Cobbett, 1991; Behm€uller et al., 2016).A study showed that two members of UDP-galactose (UDP-Gal)epimerases (UGE1 and 3) displayed both UDP-Xyl 4-epimeraseand UDP-Glc 4-epimerase activities in vitro (Kotake et al., 2009).Unlike the mur4 mutant, however, the uge1 uge3 double mutantdoes not contain a reduced Ara content (R€osti et al., 2007),although AGP synthesis is compromised (Seifert et al., 2002).

Previous studies have shown that high sugar response 8 (hsr8),an allele of the mur4 mutant, displays glucose-hypersensitivehypocotyl elongation and development under dark conditions(Li et al., 2007; Seguela-Arnaud et al., 2015). Li et al. (2007)found that the glucose-hypersensitive phenotype of the hsr8mutant can be rescued by boric acid, indicating that the alteredcell wall is the cause of the glucose-hypersensitive phenotype ofhsr8 mutant plants. Boric acid also restores wild-type stature andstem tensile strength in the low cell wall fucose mutant murus1,implicating RG-II�borate complexes as contributors to cellintegrity (O’Neill et al., 2001; Ryden et al., 2003). Genetic stud-ies have indicated that the mutations in PRL1 (PleiotropicRegulatory Locus 1) or the MED25 and MED8 subunits of theMediator complex suppress the glucose-hypersensitive phenotypeof hsr8 mutant (Li et al., 2007; Seguela-Arnaud et al., 2015).

In this study, we performed a genetic screen for mutants thatare hypersensitive to salt stress in Arabidopsis and identified sev-eral mutants that showed abnormal root elongation under saltstress. One mutant, in which the UDP-D-xylose 4-epimerase1(UXE1) gene (MUR4) is mutated, was characterised in this study.Our results showed that mutations in the MUR4 gene, which isrequired for the biosynthesis of UDP-Ara, led to a reduced rootelongation under salt stress, a phenotype that could be recoveredby applying exogenous Ara. In addition, we demonstrated thatthe three paralogues of MUR4 also contributed to the biosynthe-sis of UDP-Ara and are essential for the maintenance of rootelongation in Arabidopsis.

Materials and Methods

Plant materials and growth conditions

All Arabidopsis plants in this study were on the Columbiabackground. mur4-1 (CS8568), hsr8-2 (SALK_010548), dur

(SALK_143736), murl (CS877404), mee25 (SALK_025826),prl1 (SALK_039427), aba2-1, and abi1-1 mutants were obtainedfrom Arabidopsis Biological Resource Center (ABRC, Ohio StateUniversity, Columbus, OH). ara1-2 ara2-1 double mutant andhpgt1-1 hpgt2-1 hpgt3-1 triple mutant have been described previ-ously (Ogawa-Ohnishi & Matsubayashi, 2015; Behm€uller et al.,2016). med25 hsr8-1, med8 hsr8-1, and mur4-1 prl1-1 were pro-vided by Dr Michael W. Bevan (Li et al., 2007; Seguela-Arnaudet al., 2015). Plants were grown at 23°C with a long-day lightcycle (16 h light : 8 h dark).

Mapping by whole-genome resequencing

snrk2.1/3/4/5/6/7/8/9/10 nonuple (2W) mutant seeds (Fujii et al.,2011) were subjected to EMS mutagenesis. For whole-genomeresequencing-based mapping, the 20-2 mutant was backcrossedto 2W. In the F2 generation, c. 120 plants with slower root elon-gation under salt stress were collected and mixed together forgenomic DNA isolation using the DNeasy Plant Maxi kit (Qia-gen). Whole genomic DNAs were sequenced by Illumina. Rese-quencing reads from 2W and 20-2 were aligned to the TAIR10reference genome using BWA (Li & Durbin, 2009). SAM fileswere generated and converted to BAM files. PCR duplicationswere removed using MarkDuplicates.jar in Picard (http://broadinstitute.github.io/picard/). HaplotypeCaller in GATK (McKennaet al., 2010) was used to identify SNPs in both wild-type andmutant samples. A five-SNP window was used to calculate theratio of mutations. Based on the mutation ratio of each window,a line plot was generated along all five chromosomes. The peakposition was marked on the chromosome. The SNPs near thepeak position were selected to identify the causal mutations.

Construction of plasmids

To examine the expression pattern of MUR4, MEE25, DUR andMURL in tissues, a 2-kb fragment of the promoter sequence ofeach gene was amplified and cloned into the pMDC162 vector,which contains a b-glucuronidase reporter gene. The constructswere transformed into wild-type plants. The constructs 35S::MUR4–YFP, 35S::MURL–YFP, 35S::DUR–YFP, and 35S::MEE25–YFP were generated by amplifying the coding sequenceof each gene and cloning them into the pDONR207 ENTRYvector using BP clonase II (Life Technologies). LR clonase II wasused for the recombination of fragments to the destination vectorpEarleyGate 101. For split luciferase assay, the coding sequencesof MUR4, DUR, MURL, and MEE25 were recombined intopEarleyGate vectors that were modified from the origi-nal pCAMBIA-nLUC and pCAMBIA-cLUC vectors (Chenet al., 2008). Primers used for constructs are listed in SupportingInformation Table S1.

Microscopic observation of root morphology

To observe root morphology, 5-d-old seedlings were transferredfrom Murashige and Skoog (MS) medium to 100 mM NaClmedium and grown for an additional 2 d. The root tips were

New Phytologist (2019) � 2019 The Authors

New Phytologist� 2019 New Phytologist Trustwww.newphytologist.com

Research

NewPhytologist2

excised and the elongation zones were observed with bright fieldoptics on a Zeiss Axio Observer D1 microscope. For scanningelectron microscopy, roots were frozen in liquid nitrogen andthen sputter-coated and examined with an XL 30 FEG (Philips)cryoscanning electron microscope.

Monosaccharide analysis

The rosette leaves from 35-d-old plants or the roots of 15-d-oldseedlings were collected for monosaccharide analysis. Cell wallswere isolated as described previously (Mertz et al., 2012). In brief,the ground samples were suspended in a buffer with 50 mM Tris(pH 7.2) and 1% sodium dodecyl sulphate (SDS). After beingvortexed and centrifuged at 2500 g for 5 min, the supernatantswere removed and the pellets were resuspended in the samebuffer with and heated at 65°C for 20 min. The pellets werewashed three times with 50°C water, three times with 50%ethanol, and three times with water at ambient temperature. Cellwalls were freeze dried for 2 d before analysis of monosaccharides.About 2 mg of cell wall of each sample was hydrolysed in glassvials with 2M trifluoroacetic acid (TFA) supplemented with0.5 lM of myo-inositol as an internal standard. The samples wereincubated at 120°C for 90 min. After the samples cooled to roomtemperature, 0.5 ml of tert-butyl alcohol was added, and the sam-ples were dried overnight under a stream of air at 40°C. A 0.6-mlvolume of a solution containing 0.5 ml of fresh 20 mg ml�1

NaBH4 in dimethyl sulphoxide (DMSO) plus 0.1 ml of 1 MNH4OH was added. The samples were incubated at 45°C for90 min and then neutralised with l00 ll of glacial acetic acid. A100-ll volume of 1-methylimidazole and 0.75 ml of anhydrousacetic anhydride were added, and the mixtures were incubated at45°C for 30 min. Finally, 1.5 ml of H2O and 1 ml ofdichloromethane were added, and the samples were centrifugedat 600 g for 3 min and washed five times with H2O. The CH2Cl2was evaporated at 40–45°C, and the samples were re-dissolved in0.5 ml of CH2Cl2 for GC-MS analysis. Samples were analysedby gas�liquid chromatography on a 0.25-mm9 30-m SP-2330column (Supelco, Bellefonte, PA, USA).

Quantitative real-time RT-PCR analysis

Total RNA was extracted from whole seedlings using a plantRNA kit (Omega Bio-tek, Norcross, CA, USA) according to themanufacturer’s instructions. 2 lg RNA was treated with DNaseusing a Turbo DNA-free kit (Applied Biosystems, San Jose, CA,USA/Ambion), and was then transcribed to cDNA using reversetranscriptase enzyme M-MLV (Promega). Quantitative real-timePCR was performed using PerfeCTa SYBR Green Fastmix(Quanta Biosciences, Gaithersburg, MD, USA). Primers used forqRT-PCR are listed in Table S1.

Protein isolation and immunoblotting

Proteins were extracted from 10-d-old seedlings or tobacco leavesusing the following buffer: 50 mM Tris�HCl pH 7.5, 150 mMNaCl, 10% glycerol, 5 mM DTT, 1% IGEPAL CA-630, 1 mM

Na2MoO4.2H2O, 1 mM NaF, 1.5 mM Na3VO4, 1 mM EDTA,1 mM PMSF, 10 mM antipain, 10 mM aprotinin, and 10 mMleupeptin. Samples were incubated for 1 h and centrifuged at18 000 g and 4°C for 10 min. The supernatants were transferredto new 1.5 ml tubes, and the total protein concentration of eachsample was measured using Bradford Reagent (Bio-Rad). Anequivalent quantity of protein for each sample was loaded andseparated by SDS-PAGE. Immunoblotting was performed usinganti-GFP antibody.

Yariv staining and quantification of AGPs

After 10-d-old seedlings had been placed in half-strength MS(½MS) liquid medium for 24 h, they were placed in ½MS liquidmedium with or without 100 mM NaCl. Roots from treated anduntreated seedlings were immersed in 2 mM Yariv reagent for 1 hwith shaking and were then washed three times with distilledwater. The stained roots were examined and photographed usinga Leica EZ4HD stereomicroscopy. AGPs were quantified as pre-viously described (Gao et al., 1999; Schultz et al., 2000).

For the quantification of AGPs in growth medium, 12-d-oldwild-type seedlings were transferred to ½MS liquid medium withor without NaCl (120mM) and incubated for 6 h. After removingseedlings, the liquid media were filtered with two layers of Mira-cloth (475855-1R; Millipore) and centrifuged at 1000 g for 5 min.The media were dialysed against distilled water for 24 h and freezedried. The dried pellets were dissolved in 500 ll of 1% CaCl2.Total carbohydrate content was determined using the phenol�sul-phuric method (Dubois et al., 1956). b-D-Galactosyl�Yarivreagent was added with a final concentration of 1 : 45 of the totalcarbohydrate. The solutions were mixed and incubated at 4°C for2 h, and then centrifuged at 10 000 g for 15min. The pellets werewashed twice using 1ml 2% CaCl2. Pellets were then dissolved in500 ll of 20mM NaOH, and absorbance at wavelength 457 nmwas measured. Gum arabic in 1% (w/v) CaCl2 in water was usedas a standard with 1% CaCl2 as a blank.

GUS staining

For histochemical GUS staining, MUR4-GUS, MEE25-GUS,DUR-GUS and MURL-GUS transgenic plants were incubated ina staining buffer (10 mM EDTA, 0.5 mM K3Fe[CN],100 mMsodium phosphate, pH 7.0, 5-bromo-4-chloro-3-indolyl b-D-glucuronide, 0.5 mM K4Fe[CN], and 0.1% Triton X-100) for18 h. The stained samples were washed in 70% ethanol toremove all chlorophyll.

Cell integrity assay

Next, 6-d-old seedlings grown on ½MS medium were transferredto medium supplemented with 140 mM NaCl. After growth for12 h, the roots were stained with propidium iodide (PI) beforethey were evaluated by confocal microscopy (Zeiss, Oberkochen,Germany). Cell death was defined as fully stained by PI. Thenumber of burst cells in the elongation zone of roots were calcu-lated and at least 10 roots for each genotype were analysed.

� 2019 The Authors

New Phytologist� 2019 New Phytologist TrustNew Phytologist (2019)

www.newphytologist.com

NewPhytologist Research 3

Split luciferase assay

Agrobacterium strains (GV3101) harbouring the correspondingconstructs was grown in liquid Lysogeny Broth (LB) and harvestedby centrifugation. The pellets were resuspended in an injectionbuffer (10mM MgCl2, 10 mM MES, pH 5.6, and 100 lM ace-tosyringone) with a final concentration of OD600 = 0.5. The strainswere incubated at room temperature for 2 h before infiltration. Forco-infiltration, equal volume of two different strains carrying theindicated nLUC and cLUC constructs were mixed and infiltratedinto N. benthamiana using a 1-ml disposable syringe. After 48 h,luciferin was sprayed on the infiltrated leaves, and the fluorescencewas detected with a charge coupled device (CCD) camera (Prince-ton Instruments, Trenton, NJ, USA).

Cellular localisation and confocal microscopy

A. tumefaciens (GV3101) harbouring the GmMan1-mCherry plas-mid was mixed with bacteria expressing 35S::MUR4–YFP, 35S::MURL–YFP, 35S::DUR–YFP, or 35S::MEE25–YFP. The mixedbacteria were coinfiltrated into N. benthamiana leaves and YFPand mCherry signals were visualised after infiltration for 2 d byusing confocal laser scanning microscopy.

Results

MUR4 is required for root elongation under salt stress

To identify genes that are required for salt stress tolerance, wegenerated an EMS-mutagenised mutant pool and screened formutants that were hypersensitive to salt stress. One of themutants, designated 20-2, was defective in root elongation underhigh-salinity conditions (Fig. S1a). By using whole-genome rese-quencing-based genetic mapping, a mutation in the MUR4 genewas identified in the 20-2 mutant (Fig. S1b). To determine thatMUR4 was the mutated gene responsible for the salt-hypersensitive phenotype of the 20-2 mutant, we examined thephenotype of two published alleles of the mur4 mutant, mur4-1and hsr8-2 (Burget et al., 2003; Li et al., 2007), under high salt.Both mutant alleles showed normal root elongation on MSmedium but much slower root elongation on MS medium

supplemented with NaCl compared with the wild-type plants(Fig. 1a,b). However, the growth of the aerial part of mur4-1/hsr8-2 mutants appeared to be unaffected by salt stress (Fig. 1a).These results indicated that MUR4 is required for root elonga-tion under salt stress. Because the phenotypes of the mur4-1 andhsr8-2 mutant alleles were similar to that of the 20-2 mutant, weused these two mutant alleles for further study.

To further determine the salt-hypersensitive phenotype of themur4 mutant, we sowed the seeds of the wild-type and mur4-1on MS and NaCl media and assessed the rates of seed germina-tion and cotyledon greening. The results showed that the rates ofseed germination and cotyledon greening of the mur4-1 mutantwere normal on MS medium but were much slower than those ofthe wild-type on NaCl medium (Fig. 1c). The mur4-1 mutantsalso exhibited slower root elongation than the wild-type on othersalt media, such as KCl, NaNO3, and KNO3 media (Fig. 1d,e).To determine whether high salinity reduced the root elongationof the mur4 mutant as a consequence of ion toxicity or osmoticstress, we transferred seedlings to MS medium containing a highconcentration of mannitol, which can induce osmotic stress butnot ion toxicity. The results showed that the roots of the mur4-1/hsr8-2 mutants elongated normally on mannitol medium(Fig. 1a,b), indicating that the salt-hypersensitive phenotype ofthe mur4-1/hsr8-2 mutants was caused by ion toxicity.

To confirm that the mutated MUR4 was the genetic defectresponsible for the salt-hypersensitive phenotype of the mur4mutant, the wild-type MUR4 coding sequence driven by theCaMV 35S promoter was transformed into the mur4-1 mutant.Two independent lines with high transcript levels of the MUR4gene were selected for phenotyping under salt stress (Fig. 1f).Both of the MUR4 overexpression lines fully rescued the shortroot phenotype of the mur4-1 mutant on NaCl and KCl media(Fig. 1g,h). Notably, the roots of the transgenic plants overex-pressing the MUR4 gene elongated slightly faster than those ofthe wild-type on NaCl medium (Fig. 1g,h).

The transcript level ofMUR4 is upregulated by NaCl andABA

Because MUR4 is important for salt tolerance, we tested whethersalt stress may regulate the expression of the MUR4 gene. After

Fig. 1 MUR4 is required for root elongation under salt stress. (a) Here, 5-d-old Arabidopsis seedlings of the wild-type and two alleles of themur4mutants,mur4-1 and hsr8-2, were transferred to Murashige and Skoog (MS), MS+NaCl (100mM), and MS+mannitol (200mM) media, respectively. The seedlingswere photographed 5 d after transfer. (b) Root lengths of the wild-type,mur4-1, and hsr8-2 grown on MS, MS+NaCl (100mM), and MS+mannitol(200mM) media. Values are means� SD (n = 10); *, P < 0.05 (Student’s t-test). (c) Seed germination and cotyledon greening rates of the wild-type andmur4-1 grown on MS and MS+NaCl (100mM) media. Error bars indicate the SD of three biological replicates. (d) Here, 5-d-old seedlings of the wild-typeandmur4-1were transferred to MS, MS+KCl (100mM), MS+NaNO3 (100mM), and MS+KNO3 (100mM) media. The seedling were photographed 5 dafter transfer. (e) Root lengths of the wild-type andmur4-1 seedlings grown on MS, MS+KCl (100mM), MS+NaNO3 (100mM), and MS+KNO3

(100mM) media. Values are means� SD (n = 10); *, P < 0.05 (Student’s t-test). (f) qRT-PCR analysis of the transcript level ofMUR4 in the wild-type,mur4-1, and two independentMUR4 overexpressing lines in themur4-1mutant background (mur4-1/MUR4OX1 andmur4-1/MUR4OX2). ACTIN8 wasused as the internal control. Values are means� SD (n = 3); *, P < 0.05 (Student’s t-test). (g) Phenotypes of the wild-type,mur4-1, andMUR4

overexpressing plants grown on MS, MS+NaCl (100mM), and MS+KCl (100mM) media. (h) Root lengths of seedlings grown on MS, MS+NaCl(100mM), and MS+KCl (100mM) media. Values are means� SD (n = 10); *, P < 0.05 (Student’s t-test). (i, j) qRT-PCR analysis of the transcript level ofMUR4 in the wild-type plants after treatment with NaCl (150mM) (i) or ABA (50 lM) (j). ACTIN8 was used as the internal control. Values are means� SD(n = 3); *, P < 0.05 (Student’s t-test). (k) qRT-PCR analysis of the transcript level ofMUR4 in the wild-type, aba2-1, and abi1-1 before and after NaCl(150mM) treatment. Values are means� SD (n = 3); *, P < 0.05 (Student’s t-test). (l)MUR4-GFP transgenic plants were treated with NaCl (150mM) for0, 1, 3 or 6 h. Immunoblottings were performed using anti-GFP and anti-actin antibodies.

New Phytologist (2019) � 2019 The Authors

New Phytologist� 2019 New Phytologist Trustwww.newphytologist.com

Research

NewPhytologist4

wild-type seedlings were treated with NaCl for 0, 1, 3, and 6 h,quantitative real-time (qRT)-PCR analysis was performed. Theresult showed that the expression of MUR4 gene was upregulatedafter NaCl treatment (Fig. 1i). Similarly, treatment of wild-type

seedlings with ABA also increased the expression of MUR4 gene(Fig. 1j), which prompted us to test whether the salt-upregulationexpression of MUR4 gene expression depends on the ABAresponse pathway. To this end, the aba2-1 mutant, in which the

WTmur4-1

020406080

100120

020406080

100120

Ger

min

atio

n ra

te (%

)

Ger

min

atio

n ra

te (%

)

MS NaCl

020406080

100120

Gre

en c

otyl

edon

(%)

020406080

100120

Gre

en c

otyl

edon

(%)

WT mur4-1 hsr8-2 WT mur4-1 hsr8-2 WT mur4-1 hsr8-2

012345678

MS NaCl Mannitol

WTmur4-1hsr8-2

Roo

t len

gth

(cm

)

MS NaCl Mannitol

MS NaNO3KCl KNO3

01234567

MS KCl NaNO3 KNO3

Roo

t len

gth

(cm

) WTmur4-1

MS NaCl

WT mur4-1 WT mur4-1 WT mur4-1 WT mur4-1

(a) (b)

(c)

(d) (e)0 12 24 36 48 60 72 84 96 108

120

132

1440 12 24 36 48 60 72 84 96 1 08

1200 12 24 36 48 60 72 84 96 108

120 0 12 24 36 48 60 72 84 96 108

120

1 32

144

WTmur4-1

WTmur4-1

WTmur4-1

* *

* * *

(f) (g)

MS NaCl KCl

00.010.020.030.040.050.06

MU

R4/

ACTI

N

0123456789

Roo

t len

gth

(cm

)

MS NaCl KCl

(h)

*

*

**

00.0050.0100.0150.0200.0250.030

0 h 1 h 3 h 6 h

NaCl treatment

MU

R4/

AC

TIN 0 1 3 6

anti-GFP

anti-Actin

NaCl treatment00.0050.0100.0150.0200.0250.0300.035

WTaba2-1abi1-1

MU

R4 /

AC

TIN

0 h 6 h0

0.0050.0100.0150.0200.0250.030

MU

R4/

AC

TIN

0 h 1 h 3 h 6 h

ABA treatment NaCl treatment MUR4-GFP

h

(i) (j) (k) (l)

* *

**

*

* *

mur4-1

WT

WT mur4-1mur4-1/MUR4ox1mur4-1/MUR4ox2

(h) (h) (h) (h)

mur4-1/

MUR4ox1

mur4-1/

MUR4ox2

mur4-1

WT mur4

-1/

MUR4ox1

mur4-1/

MUR4ox2

mur4-1

WT mur4

-1/

MUR4ox1

mur4-1/

MUR4ox2

mur4-1W

T

mur4-1/

MUR4ox1

mur4-1/

MUR4ox2

� 2019 The Authors

New Phytologist� 2019 New Phytologist TrustNew Phytologist (2019)

www.newphytologist.com

NewPhytologist Research 5

ABA biosynthesis is defective (Gonz�alez-Guzm�an et al., 2002),and the abi1-1 mutant, in which the ABA core signaling pathwayis repressed (Gosti et al., 1999), were treated with NaCl and thetranscript level of MUR4 gene was examined in these twomutants. The upregulation of MUR4 expression by salt stress wasattenuated in both aba2-1 and abi1-1 mutants (Fig. 1k). Consis-tent with the increased transcript level of MUR4 gene in thewild-type plants after salt treatment, the protein level of MUR4was also increased after salt treatment (Fig. 1l).

Exogenous Ara rescues the salt-hypersensitive phenotypeof themur4mutant

Consistent with previous studies (Burget & Reiter, 1999; Liet al., 2007), our results showed that the Ara content in the cellwalls of rosette leaves was reduced by c. 50% in both the mur4-1and hsr8-2 mutant alleles, whereas transgenic plants overexpress-ing MUR4 had a higher Ara content than the wild-type plants(Fig. 2a). These results confirmed that MUR4 is required for thebiosynthesis of UDP-Ara in Arabidopsis. We also examined theAra content in 15-d-old seedlings before and after salt treatment,and found that salt treatment slightly reduced the Ara content inthe wild-type, but substantially reduced the Ara content in boththe mur4-1 and hsr8-2 mutant alleles (Fig. 2b). To determinewhether the reduced root elongation of mur4-1/hsr8-2 mutantsunder salt stress was caused by the reduced level of Ara, we trans-ferred seedlings to NaCl medium supplemented with exogenousAra. We found that the root elongation of mur4-1/hsr8-2mutants on NaCl medium was restored to the wild-type level byexogenous Ara (Fig. 2c,d). By contrast, supplementation withtwo other monosaccharides, Glc and Xyl, failed to rescue the salt-hypersensitive phenotype of the mur4-1/hsr8-2 mutants (Fig. 2c,d). Together, these results indicated that Ara biosynthesis isimportant for maintaining root elongation under salt stress.

ARA1 and ARA2 catalyse the first step in returning Ara to thenucleotide sugar pool via a salvage pathway (Dolezal & Cobbett,1991; Behm€uller et al., 2016). To understand the role of the sal-vage pathway in salt stress tolerance, we assessed the phenotypesof ara1-2 ara2-1 double and hsr8-2 ara1-2 ara2-1 triple mutantson NaCl medium. The ara1-2 ara2-1 double mutants showed asimilar root length as the wild-type, and the hsr8-2 ara1-2 ara2-1triple mutants exhibited a similar root length as the hsr8-2 singlemutant under salt stress (Fig. 2e,f). Further study showed thatexogenous Ara recovered the short root phenotype of the hsr8-2single mutant, but not the short root phenotype of the hsr8-2ara1-2 ara2-1 triple mutant under salt stress (Fig. 2e,f), which

indicates that hsr8-2 plants required the ARA1/ARA2-mediatedsalvage pathway to be restored to wild-type by exogenous Ara.

Previous studies have shown that mutation of PRL1,MED25, or MED8 suppresses the sugar-hypersensitive hypoc-otyl elongation phenotype of the mur4-1/hsr8-1 mutant underdark conditions (Li et al., 2007; Seguela-Arnaud et al., 2015).Here, we examined the phenotype of med25 hsr8-1, med8hsr8-1, and mur4-1 prl1-1 on salt medium, and found thatthe roots of these mutants were only slightly longer thanthose of mur4-1/hsr8-1 (Fig. S2a,b), indicating that MUR4/HSR8-mediated salt stress response pathway is distinct fromthe sugar-response pathway. Ara is involved in the decorationof side chains B and D of cell wall RG-II (Bar-Peled et al.,2012). Here, we found that application of boric acid, whichis essential for the cross-linking of RG-II (O’Neill et al.,2001), partially suppressed the short root phenotype of thehsr8 mutant under salt stress (Fig. 2g), indicating that RG-IIalso contributes to root growth under salt stress.

Cell�cell adhesion is disrupted in themur4mutant undersalt stress

Because the mur4 mutant showed abnormal root growth undersalt stress, we used brightfield microscopy to examine the mor-phology of mur4-1 root cells before and after NaCl treatment.Without NaCl treatment, root cell morphology did not differbetween the wild-type and mur4-1 mutant. Under salt stress,however, the wild-type plants were able to maintain straight rootelongation, whereas the root cells of mur4-1 mutant were twistedand their cell�cell adhesion was severely disrupted (Fig. 3a). The35S::MUR4 plants of the mur4 mutant had normal cell elonga-tion and cell–cell adhesion (Fig. 3a). Using scanning electronmicroscopy, we confirmed that the root cells of the mur4-1mutant under salt stress had abnormal shapes and were separatedby large gaps (Fig. 3b). In many cases, the epidermal cells of themur4-1 mutant were peeled away from the roots under highsalinity (Fig. 3b). These results indicated that the Ara-containingcomponents are critical for the maintenance of cell wall integrityand cell–cell adhesion under salt stress. In addition, we foundthat application of boric acid partially rescued the defective cell–cell adhesion of mur4-1 mutant under high salinity (Fig. S3a).Furthermore, we found that the elongation cell number in theroots of the mur4 mutant was not affected under normal condi-tions but was substantially reduced under high salinity comparedwith that in the wild-type (Fig. 3c). Plants with disrupted cellwall integrity showed increased cell bursting under high salinity

Fig. 2 Complementation of the salt-hypersensitive phenotype ofmur4mutant by exogenous Ara (a) Quantitative assessment of Ara content in theArabidopsis rosette leaves of the wild-type,mur4-1, hsr8-2, andMUR4 overexpressing plants (mur4-1/MUR4OX1). Values are means� SD (n = 3); *,P < 0.05 (Student’s t-test). (b) Quantitative measurement of Ara content in the Arabidopsis seedlings of the wild-type,mur4-1, and hsr8-2 before and afterNaCl treatment. Values are means� SD (n = 3); *, P < 0.05 (Student’s t-test). (c) Phenotypes of the wild-type,mur4-1 and hsr8-2 seedlings grown on MS,MS+NaCl (100mM), MS+NaCl+Ara (20mM), MS+NaCl+Glc (20mM), MS+NaCl+Xyl (20mM), and MS+Ara (20mM) media. (d) Root lengths of eachgenotype grown on MS, MS+NaCl (100mM), MS+NaCl+Ara (20mM), MS+NaCl+Glc (20mM), and MS+NaCl+Xyl (20mM) media. Values aremeans� SD (n = 10); *, P < 0.05 (Student’s t-test). (e) Phenotypes of the wild-type, hsr8-2, ara1-2 ara2-1, and hsr8-2 ara1-2 ara2-1 on MS, MS+NaCl(100mM), and MS+NaCl+Ara (20mM) media. (f) Root lengths of the wild-type, hsr8-2, ara1-2 ara2-1, and hsr8-2 ara1-2 ara2-1 on MS, MS+NaCl(100mM), and MS+NaCl+Ara (20mM) media. Values are means� SD (n = 10); *, P < 0.05 (Student’s t-test). (g) Root lengths of the wild-type and hsr8-2seedlings grown on MS, MS+NaCl (100mM), and MS+NaCl+boric acid (2 mM) media. Values are means� SD (n = 10); *, P < 0.05 (Student’s t-test).

New Phytologist (2019) � 2019 The Authors

New Phytologist� 2019 New Phytologist Trustwww.newphytologist.com

Research

NewPhytologist6

(Feng et al., 2018; Sechet et al., 2018). Our result showed thatthe mur4 mutants exhibited more burst cells under high salinitycompared with the wild-type (Figs 3d, S4), suggesting that arabi-nose-containing polymers are required for the maintenance ofcell wall integrity under high salinity.

The short root phenotype of themur4mutant under saltstress is caused by the reduced levels of AGPs

The carbohydrate components of AGPs are branched galactansheavily decorated with Ara residues (Knoch et al., 2014). To

(a)

MS NaCl NaCl+20 mM arabinose

NaCl+20 mM glucose NaCl+20 mM xylose MS+20 mM arabinose

mur4-1WT hsr8-2 mur4-1WT hsr8-2 mur4-1WT hsr8-2

mur4-1WT hsr8-2 mur4-1WT hsr8-2 mur4-1WT hsr8-2

(c)

(b)

02468

101214

WT

mur4-1

hsr8-

2

Arab

inos

e co

nten

t

0

5

10

15

20

25

Arab

inos

e co

nten

t

(d)

in ro

sette

leav

es (%

)in

see

dlin

gs (%

)

* *

*

MS NaCl

WT mur4-1 hsr8-2

01234567

MS NaCl NaCl+

WT hsr8-2

Roo

t len

gth

(cm

)

boric acid

(g)

0123456789

MS NaCl NaCl+arabinose

NaCl+glucose

NaCl+xylose

WT mur4-1 hsr8-2

Roo

t len

gth

(cm

)

** * * * *

WT hs

r8-2ara

1/2hs

r8-2 a

ra1/2

WT hs

r8-2

ara1/2

hsr8-

2 ara1

/2

WT hs

r8-2

ara1/2

hsr8-

2 ara1

/2

MS NaCl NaCl+arabinose

0123456789

MS NaCl

WThsr8-2

ara1-2 ara2-1hsr8-2 ara1-2 ara2-1

Roo

t len

gth

(cm

)

(e)

(f)

* * *

* *

* *

*

NaCl+arabinose

mur4-1/

MUR4ox1

� 2019 The Authors

New Phytologist� 2019 New Phytologist TrustNew Phytologist (2019)

www.newphytologist.com

NewPhytologist Research 7

determine whether the reduced content of Ara in the mur4mutant affected the formation of AGPs, we grew wild-type andmur4-1 seedlings on MS and NaCl media and stained the rootswith Yariv reagent, which specifically binds to AGPs (Wordenet al., 2012; Kitazawa et al., 2013). In the roots of the wild-typeseedlings, the level of AGPs was slightly reduced by NaCl treat-ment (Fig. 4a). In the roots of the mur4-1 mutant, the level ofAGPs was slightly reduced under normal condition, but wasmarkedly decreased under high salinity compared with the wild-type (Fig. 4a). Quantification of AGPs using high-performanceliquid chromatography (HPLC)�mass spectrometry alsoshowed that the AGPs were substantially decreased in the mur4-1 mutants after NaCl treatment compared with the wild-type(Fig. 4b). To test whether AGPs are released from cells underhigh salinity, we measured AGP content in the growth mediumafter salt treatment. Our results indicated that high salinityaccelerated the release of AGPs to growth medium in Arabidop-sis (Fig. 4c). To understand whether the reduction of AGPs wasthe cause of the reduced root elongation of the mur4 mutantunder salt stress, we examined the phenotype of mur4-1/hsr8-2mutants on NaCl medium supplemented with gum arabic, a

commercial source of Acacia AGPs (Kitazawa et al., 2013). Theresults showed that the exogenous gum arabic restored the rootelongation of mur4-1/hsr8-2 mutants under salt stress to that ofthe wild-type level (Fig. 4d,e). Gum arabic also restored nearwild-type level growth in the hsr8-2 ara1-2 ara2-1 triple mutant(Fig. 4f), which lacked the Ara salvage pathway, demonstratingthat the restoration was by AGPs and rather than by Arahydrolysed from the AGPs by endogenous cell wall enzymes.The disrupted cell–cell adhesion of the mur4-1 mutant underhigh salinity was also recovered when exogenous gum arabicwas supplied (Fig. S3b).

Previous studies have shown that HPGT1, HPGT2, andHPGT3 are required for the galactosylation of AGPs and muta-tion of the genes encoding these three proteins greatly reducesAGP content (Ogawa-Ohnishi & Matsubayashi, 2015). To fur-ther determine that AGPs are required for root elongation undersalt stress, we assessed the phenotype of the hpgt123 triple mutantgrown on salt medium. Like the mur4 mutant, the hpgt123mutant had normal root elongation on MS medium but showedreduced root elongation under salt stress (Fig. 4g,h). The defectin root elongation of the hpgt123 mutant under high salinity was

MS NaCl MS NaCl MS NaCl

WT mur4-1 mur4-1/MUR4-YFP

WT WT mur4-1mur4-1

MS NaCl MS NaCl

(a)

(b)

05

1015202530

MS NaCl

WT mur4-1

Num

ber o

f elo

ngat

ion

cells

*

02468

101214

Num

ber o

f bur

st c

ells

WT mur4-1

(c) (d)

*

Fig. 3 Cell–cell adhesion is disrupted in themur4mutant under salt stress in Arabidopsis.(a) Morphology of root cells of the wild-type,mur4-1, andMUR4 overexpressing seedlingsbefore and after NaCl (100mM) treatmentfor 24 h was imaged by light microscopy.Bar, 50 lm. (b) Morphology of root cells ofthe wild-type andmur4-1 before and afterNaCl (100mM) treatment for 24 h wasdetected by using scanning electronmicroscopy. The epidermal cell that waspeeled away from the root is indicated by anarrow. Bar, 20 lm. (c) Elongation cell numberin the roots of the wild-type andmur4-1

before and after NaCl (100mM) treatment.Values are means� SD (n = 10); *, P < 0.05(Student’s t-test). (d) The number of burstcells in the root tips of the wild-type andmur4-1 after NaCl (140mM) treatment for12 h. Values are means� SD (n = 10); *,P < 0.05 (Student’s t-test).

New Phytologist (2019) � 2019 The Authors

New Phytologist� 2019 New Phytologist Trustwww.newphytologist.com

Research

NewPhytologist8

WT mur4-1 mur4-1/OX1

MS NaCl MS NaCl MS NaCl

WT hs

r8-2mur4

-1

WT hs

r8-2mur4

-1

WT hs

r8-2mur4

-1

MS NaCl NaCl + gum arabic

0

50

100

150

200

250

AG

P co

nten

t (μg

g–1

FW

)

WT mur4-1

MS NaCl

hpgt123WT hpgt123WT hpgt123WT

MS NaCl NaCl +01234567

Roo

t len

gth

(cm

)

MS NaCl NaCl +gum arabic

WT hsr8-2mur4-1

01234567

MS NaCl NaCl +gum arabic

WT hpgt123

Roo

t len

gth

(cm

)

(a) (b)

(d)

(e)

(f)

(g)

*

* * *

MS NaCl NaCl +

WT

hsr8-

2 ara1

/2

WT

hsr8-

2 ara1

/2

WT

hsr8-

2 ara1

/2

gum arabic

gum arabic

(h)

0

0.05

0.10

0.15

0.20

0.25

0.30

AG

Ps

in g

row

th

med

ium

(mg

ml–1

)

MS NaCl

(c)

*

Fig. 4 The salt-hypersensitive phenotype of themur4mutant is caused by the reduced levels of AGPs in Arabidopsis. (a) Roots of the wild-type,mur4-1,andMUR4 overexpressing plants (in themur4-1mutant background) that were treated or not treated with NaCl (100mM) were stained with Yarivreagent. Roots were photographed with a light microscope. (b) Quantification of the content of AGPs in the roots of the wild-type andmur4-1mutantbefore and after NaCl (100mM) treatment. Values are means� SD (n = 3); *, P < 0.05 (Student’s t-test). (c) Here, 12-d-old wild-type seedlings weretransferred to ½MS liquid medium with or without NaCl (120mM). After growth for 6 h, the released AGPs in the growth media were measured. Valuesare means� SD (n = 3); *, P < 0.05 (Student’s t-test). (d) Phenotypes of the wild-type, hsr8-2 andmur4-1 grown on MS, MS+NaCl (100mM), andMS+NaCl+gum arabic (1.5%) media. (e) Root lengths of the wild-type, hsr8-2 andmur4-1 seedlings grown on MS, MS+NaCl (100mM), andMS+NaCl+gum arabic (1.5%) media. Values are means� SD (n = 10); *, P < 0.05 (Student’s t-test). (f) Phenotypes of the wild-type and hsr8-2 ara1-2ara2-1 seedlings grown on MS, MS+NaCl (100 mM), and MS+NaCl+gum arabic (1.5%) media. (g) Phenotypes of the wild-type and hpgt1 hpgt2 hpgt3

(hpgt123) seedlings grown on MS, MS+NaCl (100mM), and MS+NaCl+gum arabic (1.5%) media. (h) Root lengths of the wild-type and hpgt123 grownon MS, MS+NaCl (100mM), and MS+NaCl+gum arabic (1.5%) media. Values are means� SD (n = 10); *, P < 0.05 (Student’s t-test).

� 2019 The Authors

New Phytologist� 2019 New Phytologist TrustNew Phytologist (2019)

www.newphytologist.com

NewPhytologist Research 9

rescued by the addition of exogenous gum arabic (Fig. 4g,h).These results indicated that AGPs are required for root elonga-tion under salt stress, and that the decreased level of AGPs is themain cause of the salt-hypersensitive phenotype of the mur4-1/hsr8-2 mutants.

MUR4 forms either a homodimer or heterodimers with itsisoforms

To identify proteins that are associated with MUR4 in the reg-ulation of UDP-Ara biosynthesis, we performed immunopre-cipitation�mass spectrometry (IP-MS) analysis using MUR4–GFP transgenic plants. Transgenic plants expressing free GFPwere used as the control. The peptides that belonged to theDUR protein were identified in the IP-MS data generated fromthe MUR4–GFP sample, but not from the control sample(Table S2). The DUR gene encodes an isoform of MUR4 inArabidopsis. The potential interaction of MUR4 with DURsuggests the possibility that MUR4 might form a homodimerwith itself and heterodimers with its isoforms. By aligning theMUR4 protein sequence against the Arabidopsis database atThe Arabidopsis Information Resource (TAIR), we identifiedanother two isoforms of the MUR4 protein, MEE25 andAT4G20460. Because the protein sequence of AT4G20460 wassimilar to that of MUR4, we named this gene MURL (MUR4-like). The MUR4 protein shares 79% and 82% identity withDUR and MURL, respectively. According to the proteindatabase at TAIR, the MEE25 gene encodes two protein iso-forms, and the longer isoform of MEE25 shares 88% identitywith MUR4. Split luciferase (split-LUC) assays indicated thatMUR4 interacted with itself and also with DUR and MURLbut not with MEE25 (Fig. 5a). The formation of the homod-imer of MUR4 was also supported by co-immunoprecipitation(Co-IP) assays (Fig. 5b). Strikingly, none of these three iso-forms could form homodimers or heterodimers with each other(Fig. S5).

Consistent with a previous study (Burget et al., 2003), wefound that MUR4 was localised in the Golgi when it was tran-siently expressed in Nicotiana benthamiana or stably expressed inArabidopsis (Fig. 5c,d). DUR and MURL were also localised inthe Golgi (Fig. 5c). However, the GFP fluorescence of MEE25did not overlap with but was near the fluorescence of the Golgimarker protein (Fig. 5c), suggesting that MEE25 is probablylocalised in the trans-Golgi network (TGN). The different sub-cellular localisations of MEE25 and MUR4 are consistent withthe lack of interaction between these two proteins.

MUR4 and its isoforms redundantly regulate rootelongation

LikeMUR4, all the three paralogous genes were upregulated afterNaCl and ABA treatment, and NaCl-induced expression of thesethree paralogous genes was alleviated in the aba2-1 and abi1-1mutants (Fig. S6). Histochemical staining of plants transformedwith MUR4pro:GUS, MURLpro:GUS, and DURpro:GUSrevealed that MUR4, MURL, and DUR were expressed in roots,

especially in the elongation zone, and were also expressed inleaves and flowers (Fig. S7a–c). MEE25 was highly expressed inflowers, but weakly expressed in roots and leaves (Fig. S7a–c). Toinvestigate whether the three paralogues are also involved in saltstress tolerance, we assessed the salt sensitivity of the T-DNAinsertion mutants of DUR, MURL, and MEE25 (Fig. S7d,e).The result showed that none of these was defective in root elon-gation under salt stress compared with the wild-type (Fig. S7f,g),suggesting that MUR4/HSR8 plays a major role in the regulationof root elongation under salt stress.

To investigate whether DUR, MURL, and MEE25 functionredundantly with MUR4/HSR8 in the regulation of root elonga-tion and whether these three isoforms also contribute to UDP-Arabiosynthesis, we crossed the hsr8-2 mutant allele with dur, murl,and mee25 to generate double, triple, and quadruple mutants. Theroot growth of the hsr8-2 dur, hsr8-2 mee25, and hsr8-2 dur mee25mutants on MS medium was similar to that of the wild-type(Fig. 6a,b), and the murl dur mee25 triple mutant showed slightlyreduced root elongation on MS medium compared with the wild-type. However, the hsr8-2 murl double mutant exhibited substan-tially slower root elongation on MS medium, and the root elonga-tion was further inhibited in the hsr8-2 murl mee25 and hsr8-2murl dur tripe mutants and in the hsr8-2 murl dur mee25 quadru-ple mutant (Fig. 6a,b). Under salt stress, the hsr8-2 dur and hsr8-2mee25 double mutants displayed a similar root length as the hsr8-2single mutant, while the root elongation was slightly reduced in thehsr8-2 murl, hsr8-2 dur mee25, and hsr8-2 murl mee25 mutants,but was substantially reduced in the hsr8-2 murl dur triple mutantcompared with the hsr8-2 single mutant (Fig. 6a,b). For the hsr8-2murl dur mee25 quadruple mutant, the root growth was almostcompletely arrested and the leaves were chlorotic under salt stress(Fig. 6a,b). These results showed that MUR4/HSR8 functionsredundantly with its paralogous proteins, especially with MURL,in the regulation of root elongation, and also suggested that MUR4and together with its three paralogous proteins are required for theregulation of salt stress tolerance in leaves. The short root pheno-type of hsr8-2 murl, hsr8-2 murl dur, and hsr8-2 murl dur mee25grown on MS medium was restored to that of the wild-type byadding Ara and was partially suppressed by adding gum arabic tothe medium (Fig. 6c,d), which indicated that the short roots of thehigher order mutants under normal conditions were caused byreduced levels of Ara.

MUR4 and its isoforms are redundant in the biosynthesis ofUDP-Ara

As mur4-1/hsr8-2 mutants showed only a 50% reduction in Aracontent in the rosette leaves, we investigated whether the threeisoforms of MUR4 contribute to the biosynthesis of the remain-ing Ara content. To answer this question, we measured the Aracontent in the rosette leaves of single, double, triple, and quadru-ple mutants. While the Ara content was reduced by 50% in thehsr8-2 mutant, none of the dur, murl, and mee25 single mutantsshowed reduced Ara content (Fig. S7h). The dur murl mee25triple mutant had a 14% reduction in Ara content compared withthe wild-type (Fig. 6e), indicating that these three paralogues are

New Phytologist (2019) � 2019 The Authors

New Phytologist� 2019 New Phytologist Trustwww.newphytologist.com

Research

NewPhytologist10

also involved in the biosynthesis of UDP-Ara. Unexpectedly, forhsr8-2 dur, hsr8-2 mee25, hsr8-2 murl, hsr8-2 dur mee25, hsr8-2murl mee25, hsr8-2 murl dur, and hsr8-2 murl dur mee25mutants, all of which included the hsr8-2 mutation background,their Ara contents were the same or only slightly reduced

compared with that in the hsr8-2 single mutant (Fig. 6e). Theseresults suggested that MUR4/HSR8 contributed to the majorityof Ara biosynthesis in the rosette leaves.

Because the obvious phenotypic characteristic of the mur4-1/hsr8-2 single mutant and its higher order mutants were observed

MUR4-GFP MEMB12-mCherry Bright Merge

MUR4-YFP GmMan1-Cherry Merge

DUR-YFP GmMan1-Cherry Merge

MEE25-YFP GmMan1-Cherry Merge

MURL-YFP GmMan1-Cherry Merge

1. nLUC-MUR4+cLUC-MUR42. cLUC-MUR4+nLUC-MUR43. nLUC-MUR4+cLUC4. nLUC-MUR4+cLUC5. cLUC-MUR4+nLUC6. cLUC-MUR4+nLUC

1. nLUC-MUR4+cLUC-DUR2. cLUC-DUR+nLUC-MUR43. nLUC-MUR4+cLUC4. nLUC-DUR+cLUC5. cLUC-MUR4+nLUC6. cLUC-DUR+nLUC

1. nLUC-MUR4+cLUC-MURL2. cLUC-MURL+nLUC-MUR43. nLUC-MUR4+cLUC4. nLUC-MURL+cLUC5. cLUC-MUR4+nLUC6. cLUC-MURL+nLUC

1. nLUC-MUR4+cLUC-MEE252. cLUC-MEE25+nLUC-MUR43. nLUC-MUR4+cLUC4. nLUC-MEE25+cLUC5. cLUC-MUR4+nLUC6. cLUC-MEE25+nLUC

1 2

3 4

5 6

1 2

3 4

5 6

1 2

3 4

5 6

1 2

3 4

5 6

(a)

(b)

(c)

(d)

Input

Input

IP

Co-IP

Anti-GFP

Anti-Myc

Anti-GFP

Anti-Myc

MUR4-GFPMUR4-Myc + +

+–

Fig. 5 MUR4 forms either a homodimer or heterodimers with its isoforms in Arabidopsis. (a) Split luciferase complementation assays showing that MUR4interacts with itself and with two paralogous proteins, MURL and DUR, but not with MEE25. The constructs to express the indicated fusion proteins weretransformed into Nicotiana benthamiana leaves through Agrobacterium infiltration. Luciferase activity was determined at 48 h after infiltration. nLUCrepresents N-terminal fragment of firefly luciferase; cLUC represents C-terminal fragment of firefly luciferase. (b) Co-immunoprecipitation assay showingthat MUR4 forms a homodimer in vivo.MUR4-Myc andMUR4-GFPwere expressed in N. benthamiana. Immunoprecipitation was performed by usinganti-GFP antibodies. Immunoblottings were conducted by using anti-GFP and anti-Myc antibodies. (c)MUR4-YFP, DUR-YFP,MURL-YFP, orMEE25-YFP

was coinfiltrated with GmMan1-mCherry (Golgi Marker) into N. benthamiana. GFP and cherry fluorescence were detected by using confocal microscopy.Bars, 20 lm. (d)MUR4-GFP transgenic plants were crossed withMEMB12-mCherry (Golgi marker) plants to generate stableMUR4-GFP/MEMB12-mCherry transgenic plants. GFP and cherry fluorescence were detected by using confocal microscopy. Bar, 20 lm.

� 2019 The Authors

New Phytologist� 2019 New Phytologist TrustNew Phytologist (2019)

www.newphytologist.com

NewPhytologist Research 11

in roots, we then measured the Ara content in roots. Unlike the50% reduction of Ara content in the rosette leaves, the Ara con-tent was reduced by only c. 15% in the roots of the mur4-1/hsr8-2 mutants compared with the wild-type (Fig. 6f). Under saltstress, the Ara content was reduced by c. 33% in the roots of themur4-1/hsr8-2 mutants compared with that in the wild-type

(Fig. 6f). Interestingly, the Ara content was substantially lower inthe roots of the hsr8-2 murl double mutant than that in the rootsof the hsr8-2 single mutant and was even lower in the roots of thehsr8-2 murl mee25 and hsr8-2 murl dur triple mutants (Fig. 6g).In the roots of hsr8-2 murl dur mee25 quadruple mutant, the Aracontent was reduced by c. 52% compared with the wild-type and

WThs

r8-2hs

r8-2 d

ur

hsr8-

2 mee

25

hsr8-

2 dur

mee25

hsr8-

2 murl

murl du

r mee

25

hsr8-

2 murl

dur m

ee25

hsr8-

2 murl

mee

25

hsr8-

2 murl

dur

WThs

r8-2hs

r8-2 d

ur

hsr8-

2 mee

25

hsr8-

2 dur

mee25

hsr8

murl

murl du

r mee

25

hsr8-

2 murl

dur m

ee25

hsr8-

2 murl

mee

25

hsr8-

2 murl

dur

WT

hsr8-

2 murl

hsr8-

2 murl

dur m

ee25

hsr8-

2 murl

dur

WT

hsr8-

2 murl

hsr8-

2 murl

dur m

ee25

hsr8-

2 murl

dur

WT

hsr8-

2 murl

hsr8-

2 murl

dur m

ee25

hsr8-

2 murl

dur

MS MS+arabinose MS+gum arabic

0123456789

10

WT hsr8-2 hsr8-2 dur hsr8-2 mee25 hsr8-2 murlhsr8-2 dur mee25 hsr8-2 murl mee25 hsr8-2 murl durmurl dur mee25 hsr8-2 murl dur mee25

Roo

t len

gth

(cm

)

0123456789

10R

oot l

engt

h (c

m)

MS NaClMS NaCl

(a) (b)

(c) (d)

MS MS+

WT hsr8-2 hsr8-2 murlhsr8-2 murl dur mee25

a a a a ad

b

c ce

ad

b b b b b bc

e

** *

**

*

arabinoseMS+

gum arabic

0123456789

10

Ara

bino

se c

onte

nt

WT

hsr8-

2

hsr8-

2 dur

hsr8-

2 mee

25

hsr8-

2 dur

mee25

hsr8-

2 murl

murl du

r mee

25

hsr8-

2 murl

dur m

ee25

hsr8-

2 murl

mee

25

hsr8-

2 murl

dur

in ro

sette

leav

es (%

)

(e)

ac

b b b b b b b b

05

10152025303540

Ara

bino

se c

onte

ntin

root

s (%

)

(f)

a gb b b dc

e fh

05

1015202530354045

WT hsr8-2mur4-1

Ara

bino

se c

onte

ntin

root

s (%

)

MS NaClW

Ths

r8-2

hsr8-

2 dur

hsr8-

2 mee

25

hsr8-

2 dur

mee25

hsr8-

2 murl

murl du

r mee

25

hsr8-

2 murl

dur m

ee25

hsr8-

2 murl

mee

25

hsr8-

2 murl

dur

(g)

* ** *

New Phytologist (2019) � 2019 The Authors

New Phytologist� 2019 New Phytologist Trustwww.newphytologist.com

Research

NewPhytologist12

was reduced by c. 44% compared with the hsr8-2 single mutant(Fig. 6g). These results indicated that the three isoforms ofMUR4 are involved in the biosynthesis of UDP-Ara in roots,which is consistent with the short root phenotype of the higherorder mutants under normal conditions.

ARA1 and ARA2 contribute to salt stress tolerance

Our study suggested that the isoforms of MUR4 also contributedto the UDP-Ara biosynthesis pathway (Fig. 7a). To further clarifythe contribution of both de novo and salvage pathways of UDP-Ara biosynthesis to salt stress tolerance, we generated the hsr8-2murl dur mee25 ara1-2 ara2-1 sextuple mutant. The roots of thesextuple mutant were shorter than those of the hsr8-2 murl durmee25 quadruple mutant on both MS and NaCl media, and theleaves of the sextuple mutant were more chlorotic than that ofthe quadruple mutant on NaCl medium (Fig. 7b,c). These resultssuggested that ARA1 and ARA2 are also important for salt stresstolerance. The short root phenotype of the sextuple mutant oneither MS or NaCl medium could not be recovered by exogenousAra (Fig. 7b,c), further supporting that the ARA1/ARA2-mediated salvage pathway was required for the assimilation ofexogenous Ara. Our results also showed that gum arabic partiallysuppressed the salt-hypersensitive phenotype of the sextuplemutant (Fig. 7d). Quantitative analysis of Ara content revealedthat, relative to the hsr8 murl dur mee25 quadruple mutant, thesextuple mutant had a similar Ara amount in its roots and areduced Ara content in its rosette leaves (Fig. 7e,f).

Discussion

The biosynthesis of polysaccharides in the plant cell wall isdynamically regulated in response to a variety of environmentalstresses. Our study showed that the Ara content in the cell wallwas altered after salt treatment and the plants that were disruptedin Ara biosynthesis were hypersensitive to high salinity, includingNaCl, KCl, NaNO3, and KNO3, which suggest that maintainingthe arabinosylation of cell wall components is critical for plantsto adapt to adverse environmental conditions. Even though somesalts, such as KNO3, are essential plant nutrients, high concentra-tions of these salts inhibited the growth of plants. Our studyshowed that the three paralogues of MUR4 also contributed tothe biosynthesis of UDP-Ara, and provided new insights into theUDP-Ara biosynthesis network.

Ara is involved in the decoration of a number of polymersand glycoproteins in the cell wall (Ogawa-Ohnishi et al., 2013;Borassi et al., 2016; Kotake et al., 2016). Our results showedthat application of gum arabic, a commercial product of AGPsderived from A. senega (Kitazawa et al., 2013), recovered theroot growth of mur4/hsr8 mutant under salt stress, suggestingthat the short root phenotype of the mur4/hsr8 mutant undersalt stress was caused by the alteration in root AGPs due tothe limited availability of UDP-Ara. AGPs are macromoleculesthat are typified by the decoration of the protein backbonewith Gal and Ara (Seifert & Roberts, 2007; Ellis et al., 2010;Nguema-Ona et al., 2013), and carbohydrates account for90–98% of the total weight of AGPs (Knoch et al., 2014;Ogawa-Ohnishi & Matsubayashi, 2015). AGPs are cell wallcomponents that undergo rapid turnover (Showalter, 1993;Behm€uller et al., 2016). This study and others have all shownthat high salinity accelerates the release of AGPs to growthmedium and it was proposed that AGPs could serve in Na+/Ca2+ exchange. (Lamport et al., 2006, 2014; Olmos et al.,2017). Based on these results, we hypothesised that turnover ofAGPs under salt stress may reduce Na+ activity and thereforeattenuate ion toxicity in plants. While AGPs are released underhigh salinity, new AGPs are secreted to the cell wall to main-tain cell wall integrity and root elongation. Production of newAGP protein backbones, Ara, and Gal is coordinately regulatedto generate functional AGPs before they are secreted into thecell wall (Showalter & Basu, 2016). Elevation of the transcriptand protein levels of MUR4 in the wild-type under salt stressmay counteract the reduction of AGPs. However, disruptionof Ara biosynthesis in the mur4 mutant would affect the for-mation of functional AGPs, and finally inhibit root elongationunder salt stress. These results suggested that the salt-hypersensitive phenotype of the mur4 mutant was due to thedeficiency in the structure of cell wall, but not caused by thedisrupted physiological and molecular responses within cellsunder high salinity. Cell bursting assay indicated that cell wallintegrity was disrupted in the mur4 mutant under salt stress,and resulted in a swelling phenotype and subsequently celldeath.

The role of AGPs in root cell elongation has been reported inseveral studies. Treatment of Arabidopsis seedlings with Yarivreagent, which specifically binds to AGPs, inhibited root elonga-tion (Ding & Zhu, 1997). UGE4/REB1, which encodes a UDP-D-Glc 4-epimerase, is required for the biosynthesis of UDP-D-

Fig. 6 MUR4 and its three isoforms are redundant in the regulation of root elongation and UDP-Ara biosynthesis in Arabidopsis. (a) Phenotypes of thewild-type, hsr8-2, and the higher order mutants of hsr8-2with its paralogues grown on MS and MS+NaCl (100mM) media. (b) Root lengths of the wild-type, hsr8-2, and the higher order mutants of hsr8-2 with its paralogues grown on MS and MS+NaCl (100mM) media. Values are means� SD (n = 10).For each medium, different letters represent statistically significant differences (P < 0.05, one-way analysis of variance (ANOVA)). (c) Phenotypes of thewild-type, hsr8-2 murl, hsr8-2 murl dur and hsr8-2 murl dur mee25 seedlings grown on MS, MS+Ara (20mM), and MS+gum arabic (1.5%) media. (d)Root lengths of the wild-type, hsr8-2 murl, hsr8-2 murl dur and hsr8-2 murl dur mee25 seedlings grown on MS, MS+Ara (20mM), and MS+gum arabic(1.5%) media. Values are means� SD (n = 10); *, P < 0.05 (Student’s t-test). (e) Quantification of the Ara content in the rosette leaves of each genotype.Values are means� SD (n = 3). In each panel, different letters represent statistically significant differences (P < 0.05, one-way ANOVA). (f) Quantificationof the Ara content in the roots of the wild-type, hsr8-2, andmur4-1 before and after NaCl treatment. Values are means� SD (n = 4); *, P < 0.05 (Student’st-test). (g) Quantification of the Ara content in the roots of each genotype. Values are means� SD (n = 3). In each panel, different letters representstatistically significant differences (P < 0.05, one-way ANOVA).

� 2019 The Authors

New Phytologist� 2019 New Phytologist TrustNew Phytologist (2019)

www.newphytologist.com

NewPhytologist Research 13

Gal and subsequently for the galactosylation of AGPs. The uge4/reb1 mutant is characterised by reduced root elongation and rootepidermal cell bulging (And�eme-Onzighi et al., 2002; Nguema-

Ona et al., 2006). Mutation of theMUR1 gene, which is requiredfor the biosynthesis of L-fucose, affects the fucosylation of AGPsand leads to a decrease in root cell elongation (van Hengel &

MS NaClMS+arabinose NaCl+arabinose

WT hsr8-2 murl d

ur mee25

hsr8-2 murl d

ur mee25 ara1/2

0123456789

1011

MS NaCl NaCl+arabinose

MS+arabinose

WT hsr8-2 murl dur mee25hsr8-2 murl dur mee25 ara1-2 ara2-1

Roo

t len

gth

(cm

)

05

10152025303540

Ara

bino

se c

onte

nt

0123456789

10

Ara

bino

se c

onte

nt

UDP-arabinoseUDP-xylose Arabinose-P Arabinose

MUR4MURLDUR

USPARA1ARA2

de novo pathway Salvage pathway

Cell wall

hsr8-

2ara

1/2

hsr8-

2 ara1

/2

hsr8-

2 murl du

r mee

25

hsr8-

2 murl du

r mee

25 ar

a1/2W

T

in ro

ots

(%)

in ro

sette

leav

es (%

)

(a)

(b)

(c)

(d)

WT hsr8-2 murl d

ur mee25

hsr8-2 murl d

ur mee25 ara1/2

WT hsr8-2 murl d

ur mee25

hsr8-2 murl d

ur mee25 ara1/2

WT hsr8-2 murl d

ur mee25

hsr8-2 murl d

ur mee25 ara1/2

** ** * *

ab

cd d

e

a ab b

c c

MEE25

AGPRG-II

ArabinanArabinoxylan

ExtensinArabinosyltransferases Secretion

WT

hsr8-2 murl d

ur mee25 ara1/2 WT

hsr8-2 murl d

ur mee25 ara1/2

NaCl NaCl+gum arabic (e) (f)

hsr8-

2ara

1/2

hsr8-

2 ara1

/2

hsr8-

2 murl du

r mee

25

hsr8-

2 murl du

r mee

25 ar

a1/2W

T

Fig. 7 ARA1 and ARA2 are required for salt stress tolerance in Arabidopsis. (a) Diagram showing the proteins that are involved in UDP-Ara biosynthesis viathe de novo pathway or salvage pathway. (b) Phenotypes of the wild-type, hsr8-2 murl dur mee25, and hsr8-2 murl dur mee25 ara1-2 ara2-1 on MS,MS+Ara (20mM), MS+NaCl (100mM), and MS+NaCl+Ara (20mM) media. (c) Root lengths of the wild-type, hsr8-2 murl dur mee25, and hsr8-2 murl

dur mee25 ara1-2 ara2-1 on MS, MS+Ara (20mM), MS+NaCl (100mM), and MS+NaCl+Ara (20mM) media. Values are means� SD (n = 10); *, P < 0.05(Student’s t-test). (d) Phenotypes of the wild-type and hsr8-2 murl dur mee25 ara1-2 ara2-1 seedlings grown on MS+NaCl (100mM) and MS+NaCl+gumarabic (1.5%) media. (e, f) Ara content in the rosette leaves (e) and roots (f) of each genotype. Values are means� SD (n = 3). In each panel, differentletters represent statistically significant differences (P < 0.05, one-way analysis of variance (ANOVA)).

New Phytologist (2019) � 2019 The Authors

New Phytologist� 2019 New Phytologist Trustwww.newphytologist.com

Research

NewPhytologist14

Roberts, 2002). SOS5, an AGP-like protein, is required for rootgrowth under salt stress (Shi et al., 2003). The role of AGP pro-teins in root elongation is also supported by our finding thatmutation of HPGT genes that are required for the galactosylationof AGP proteins (Ogawa-Ohnishi & Matsubayashi, 2015)showed arrested root growth under salt stress. The mechanismsunderlying the role of AGPs in root elongation are still largelyunknown. AGPs are probably required for regulation of the ori-entation of root cell elongation (Nguema-Ona et al., 2012).

Although gum arabic partially restored the wild-typegrowth in mutants of MUR4, boric acid also increased nor-mal root growth in these mutants, indicating that dimerisa-tion of RG-II also contributed to growth under NaCl stress.Four complex glycan side chains clustered in this pecticpolysaccharide form RG-II-boron diester through theirrespective apiose residues (Caffall & Mohnen, 2009). Themur1 mutation compromises the apiose-containing side chainbecause a fucosyl residue critical for structure is missing, butaddition of boric acid restores normal growth (O’Neill et al.,2001) and stem tensile strength (Ryden et al., 2003). Arapresidues are also critical subtending sugars of RG-II, and thepartial recovery of the wild-type phenotype by boric acidimplicates RG-II in growth upon challenge with NaCl.

Our study showed that Ara content in the mur4 mutant wasreduced by c. 50% in the leaves but by only 15% in the roots ofthe mur4 mutant. In the mutant of the L-fucose biosynthesis geneMUR1, the L-fucose content was reduced by more than 98% inthe cell walls of leaves but by only 60% in the cell walls of roots(van Hengel & Roberts, 2002). These results may suggest that,relative to leaves, more functionally redundant proteins con-tribute to the biosynthesis of these monosaccharides in the roots.Our results showed that the Ara content in the rosette leaves ofthe higher order mutants of HSR8 with its paralogues was similarto that in the hsr8-2 single mutant, but that the Ara content inthe roots was gradually reduced in the double, triple and quadru-ple mutants compared with the hsr8-2 single mutant, suggestingthat the three paralogous proteins of MUR4 may mainly con-tribute to the UDP-Ara biosynthesis in roots. Because Ara stillexists in the hsr8-2 murl dur mee25 quadruple mutant and evenin the hsr8-2 murl dur mee25 ara1-2 ara2-1 sextuple mutant, wespeculated that an unknown pathway may also be involved in thebiosynthesis of UDP-Ara.

Even though Ara content is reduced in the roots of mur4/hsr8mutant, the root elongation of mur4/hsr8 was similar to that ofthe wild-type on MS medium. However, hsr8-2 murl, hsr8-2murl dur, hsr8-2 murl mee25, and hsr8-2 murl dur mee25 mutantsall showed reduced root elongation under normal conditions.Because these higher order mutants contain less Ara than thehsr8-2 single mutant in roots, we speculated that a minimalthreshold of Ara content in the cell walls of the primary root isrequired to maintain normal root elongation, and that theremaining Ara content in the roots of the hsr8-2 single mutant issufficient to maintain root elongation under normal conditions.However, a substantial reduction of Ara content in the mur4/hsr8mutant under salt stress leads to the disruption of cell wallintegrity and therefore affects root elongation.

Acknowledgements

We thank Rebecca Stevenson for technical assistance, Prof.Raimund Tenhaken for providing ara1-2 ara2-1 seeds,Prof. Yoshikatsu Matsubayashi for providing hpgt123 seeds, andProf. Michael W. Bevan for providing med25 hsr8-1 and med8hsr8-1 seeds. This work was supported by the Strategic PriorityResearch Program (Grant no. XDB27040101) of the ChineseAcademy of Sciences, and monosaccharide analyses weresupported by the US Department of Energy, Office of Science,Basic Energy Sciences (Grant no. DE-SC0000997).

Author contributions

J-KZ, CZ and OZ, conceived and designed the experiments; CZand OZ performed most of the experiments; FZ, CL, LZ andYET constructed some plasmids and performed genotyping; PZ,C-CH and WAT conducted proteomics assays; NCC assisted inthe analysis of cell wall monosaccharides, and CZ and J-KZwrote the manuscript with constructive input from all authors.CZ and OZ contributed equally to this work.

ORCID

Nicholas C. Carpita https://orcid.org/0000-0003-0770-314XChuan-Chih Hsu https://orcid.org/0000-0002-7100-1401Omar Zayed https://orcid.org/0000-0003-1388-1903Ling Zhang https://orcid.org/0000-0002-2708-1176Chunzhao Zhao https://orcid.org/0000-0003-0284-2095Jian-Kang Zhu https://orcid.org/0000-0001-5134-731XPeipei Zhu https://orcid.org/0000-0003-4508-022X

References

And�eme-Onzighi C, Sivaguru M, Judy-March J, Baskin T, Driouich A. 2002.

The reb1-1mutation of Arabidopsis alters the morphology of trichoblasts, the

expression of arabinogalactan-proteins and the organization of cortical

microtubules. Planta 215: 949–958.Bar-Peled M, Urbanowicz BR, O’Neill MA. 2012. The synthesis and origin of

the pectic polysaccharide rhamnogalacturonan II-insights from nucleotide

sugar formation and diversity. Frontiers in Plant Science 3: 92.Behm€uller R, Kavkova E, D€uh S, Huber CG, Tenhaken R. 2016. The role

of arabinokinase in arabinose toxicity in plants. The Plant Journal 87:376–390.

Borassi C, Sede AR, Mecchia MA, Salgado Salter JD, Marzol E, Muschietti JP,

Estevez JM. 2016. An update on cell surface proteins containing extensin-

motifs. Journal of Experimental Botany 67: 477–487.Burget EG, Reiter W-D. 1999. The mur4mutant of Arabidopsis is partially

defective in the de novo synthesis of uridine diphospho L-arabinose. PlantPhysiology 121: 383–389.

Burget EG, Verma R, Mølhøj M, Reiter W. 2003. The biosynthesis of L-

arabinose in plants: molecular cloning and characterization of a Golgi-localized

UDP-D-xylose 4-epimerase encoded by theMUR4 gene of Arabidopsis. PlantCell 15: 523–531.

Caffall KH, Mohnen D. 2009. The structure, function, and biosynthesis of plant

cell wall pectic polysaccharides. Carbohydrate Research 344: 1879–1900.Carpita NC, Gibeaut DM. 1993. Structural models of primary cell walls in

flowering plants: consistency of molecular structure with the physical properties

of the walls during growth. The Plant Journal 3: 1–30.

� 2019 The Authors

New Phytologist� 2019 New Phytologist TrustNew Phytologist (2019)

www.newphytologist.com

NewPhytologist Research 15

Chen H, Zou Y, Shang Y, Lin H, Wang Y, Cai R, Tang X, Zhou J-M. 2008.

Firefly luciferase complementation imaging assay for protein-protein

interactions in plants. Plant Physiology 146: 368–376.Ding L, Zhu J-K. 1997. A role for arabinogalactan-proteins in root epidermal cell

expansion. Planta 203: 289–294.Dolezal O, Cobbett CS. 1991. Arabinose kinase-deficient mutant of Arabidopsisthaliana. Plant Physiology 96: 1255–1260.

Dubois M, Gilles KA, Hamilton JK, Rebers PA, Smith F. 1956. Colorimetric

method for determination of sugars and related substances. Analytical Chemistry28: 350–356.

Ellis M, Egelund J, Schultz CJ, Bacic A. 2010. Arabinogalactan-proteins: key

regulators at the cell surface? Plant Physiology 153: 403–419.Feng W, Kita D, Peaucelle A, Cartwright HN, Doan V, Duan Q, Liu M-C,

Maman J, Steinhorst L, Schmitz-Thom I et al. 2018. The FERONIA receptor

kinase maintains cell-wall integrity during salt stress through Ca2+ Signaling.

Current Biology 28: 666–675.e5.Fujii H, Verslues PE, Zhu J-K. 2011. Arabidopsis decuple mutant reveals the

importance of SnRK2 kinases in osmotic stress responses in vivo. Proceedings ofthe National Academy of Sciences, USA 108: 1717–1722.

Gao M, Kieliszewski MJ, Lamport DTA, Showalter AM. 1999. Isolation,

characterization and immunolocalization of a novel, modular tomato

arabinogalactan-protein corresponding to the LeAGP-1 gene. The Plant Journal18: 43–55.

Gonz�alez-Guzm�an M, Apostolova N, Bell�es JM, Barrero JM, Piqueras P, Ponce

MR, Micol JL, Serrano R, Rodr�ıguez PL. 2002. The short-chain alcohol

dehydrogenase ABA2 catalyzes the conversion of xanthoxin to abscisic

aldehyde. Plant Cell 14: 1833–1846.Gosti F, Beaudoin N, Serizet C, Webb AAR, Vartanian N, Giraudat J. 1999.

ABI1 protein phosphatase 2C is a negative regulator of abscisic acid signaling.

Plant Cell 11: 1897–1909.Kitazawa K, Tryfona T, Yoshimi Y, Hayashi Y, Kawauchi S, Antonov L, Tanaka

H, Takahashi T, Kaneko S, Dupree P et al. 2013. b-Galactosyl Yariv reagentbinds to the b-1,3-galactan of arabinogalactan proteins. Plant Physiology 161:1117–1126.

Klinghammer M, Tenhaken R. 2007. Genome-wide analysis of the UDP-

glucose dehydrogenase gene family in Arabidopsis, a key enzyme for

matrix polysaccharides in cell walls. Journal of Experimental Botany 58:

3609–3621.Knoch E, Dilokpimol A, Geshi N, Wallace IS. 2014. Arabinogalactan proteins:

focus on carbohydrate active enzymes. Frontiers in Plant Science 5: 198.Konishi T, Takeda T, Miyazaki Y, Ohnishi-Kameyama M, Hayashi T, O’Neill

MA, Ishii T. 2007. A plant mutase that interconverts UDP-arabinofuranose

and UDP-arabinopyranose. Glycobiology 17: 345–354.Kotake T, Hojo S, Yamaguchi D, Aohara T, Konishi T, Tsumuraya Y. 2007.

Properties and physiological functions of UDP-sugar pyrophosphorylase in

Arabidopsis. Bioscience, Biotechnology, and Biochemistry 71: 761–771.Kotake T, Takata R, Verma R, Takaba M, Yamaguchi D, Orita T, Kaneko S,

Matsuoka K, Koyama T, Reiter W-D et al. 2009. Bifunctional cytosolic UDP-

glucose 4-epimerases catalyse the interconversion between UDP-D-xylose and

UDP-L-arabinose in plants. The Biochemical Journal 424: 169–177.Kotake T, Yamaguchi D, Ohzono H, Hojo S, Kaneko S, Ishida H-K,

Tsumuraya Y. 2004. UDP-sugar pyrophosphorylase with broad substrate

specificity toward various monosaccharide 1-phosphates from pea sprouts.

Journal of Biological Chemistry 279: 45728–45736.Kotake T, Yamanashi Y, Imaizumi C, Tsumuraya Y. 2016.Metabolism of

L-arabinose in plants. Journal of Plant Research 129: 781–792.Kuang B, Zhao X, Zhou C, Zeng W, Ren J, Ebert B, Beahan CT, Deng X, Zeng

Q, Zhou G et al. 2016. Role of UDP-glucuronic acid decarboxylase in xylan

biosynthesis in Arabidopsis.Molecular Plant 9: 1119–1131.Lamport DTA, Kieliszewski MJ, Showalter AM. 2006. Salt stress upregulates

periplasmic arabinogalactan proteins: using salt stress to analyse AGP function.

New Phytologist 169: 479–492.Lamport DTA, Varnai P, Seal CE. 2014. Back to the future with the AGP-Ca2+

flux capacitor. Annals of Botany 114: 1069–1085.Li H, Durbin R. 2009. Fast and accurate short read alignment with Burrows-

Wheeler transform. Bioinformatics 25: 1754–1760.

Li Y, Smith C, Corke F, Zheng L, Merali Z, Ryden P, Derbyshire P, Waldron

K, Bevan MW. 2007. Signaling from an altered cell wall to the nucleus

mediates sugar-responsive growth and development in Arabidopsis thaliana.Plant Cell 19: 2500–2515.

Marzol E, Borassi C, Bringas M, Sede A, Rosa Rodr�ıguez Garcia D, Capece L,

Estevez JM, Garcia RD. 2018. Filling the gaps to solve the extensin puzzle.

Molecular Plant 11: 645–658.McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A,

Garimella K, Altshuler D, Gabriel S, Daly M et al. 2010. The Genome

Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA

sequencing data. Genome Research 20: 1297–1303.Mertz RA, Olek AT, Carpita NC. 2012. Alterations in cell-wall glycosyl linkage

structure of Arabidopsis murusmutants. Carbohydrate Polymers 89: 331–339.Nguema-Ona E, And�eme-Onzighi C, Aboughe-Angone S, Bardor M, Ishii T,

Lerouge P, Driouich A. 2006. The reb1-1mutation of Arabidopsis. Effect on

the structure and localization of galactose-containing cell wall polysaccharides.

Plant Physiology 140: 1406–1417.Nguema-Ona E, Coimbra S, Vicr�e-Gibouin M, Mollet J-C, Driouich A. 2012.

Arabinogalactan proteins in root and pollen-tube cells: distribution and

functional aspects. Annals of Botany 110: 383–404.Nguema-Ona E, Vicr�e-Gibouin M, Cannesan MA, Driouich A. 2013.

Arabinogalactan proteins in root-microbe interactions. Trends in Plant Science18: 1360–1385.

Ogawa-Ohnishi M, Matsubayashi Y. 2015. Identification of three potent

hydroxyproline O-galactosyltransferases in Arabidopsis. The Plant Journal 81:736–746.

Ogawa-Ohnishi M, Matsushita W, Matsubayashi Y. 2013. Identification of

three hydroxyproline O-arabinosyltransferases in Arabidopsis thaliana. NatureChemical Biology 9: 726–730.

Okamoto S, Shinohara H, Mori T, Matsubayashi Y, Kawaguchi M. 2013. Root-

derived CLE glycopeptides control nodulation by direct binding to HAR1

receptor kinase. Nature Communications 4: 1–7.Olmos E, Garc�ıa De La Garma J, Gomez-Jimenez MC, Fernandez-Garcia N.

2017. Arabinogalactan proteins are involved in salt-adaptation and vesicle

trafficking in tobacco by-2 cell cultures. Frontiers in Plant Science 8: 1092.O’Neill MA, Eberhard S, Albersheim P, Darvill AG. 2001. Requirement of

borate cross-linking of cell wall rhamnogalacturonan II for Arabidopsis growth.Science 294: 846–849.

Pe~na MJ, Darvill AG, Eberhard S, York WS, O’Neill MA. 2008.Moss and

liverwort xyloglucans contain galacturonic acid and are structurally distinct

from the xyloglucans synthesized by hornworts and vascular plants. Glycobiology18: 891–904.

Rautengarten C, Birdseye D, Pattathil S, McFarlane HE, Saez-Aguayo S,

Orellana A, Persson S, Hahn MG, Scheller HV, Heazlewood JL et al. 2017.The elaborate route for UDP-arabinose delivery into the Golgi of plants.

Proceedings of the National Academy of Sciences, USA 114: 4261–4266.R€osti J, Barton CJ, Albrecht S, Dupree P, Pauly M, Findlay K, Roberts K,

Seifert GJ. 2007. UDP-glucose 4-epimerase isoforms UGE2 and UGE4

cooperate in providing UDP-galactose for cell wall biosynthesis and growth of

Arabidopsis thaliana. Plant Cell 19: 1565–1579.Ryden P, Sugimoto-Shirasu K, Smith AC, Findlay K, Reiter W-D, McCann

MC. 2003. Tensile properties of Arabidopsis cell walls depend on both a

xyloglucan cross-linked microfibrillar network and rhamnogalacturonan II-

borate complexes. Plant Physiology 132: 1033–1040.Scheller HV, Ulvskov P. 2010.Hemicelluloses. Annual Review of Plant Biology61: 263–289.

Schultink A, Cheng K, Park YB, Cosgrove DJ, Pauly M. 2013. The

identification of two arabinosyltransferases from tomato reveals functional

equivalency of xyloglucan side chain substituents. Plant Physiology 163: 86–94.Schultink A, Liu L, Zhu L, Pauly M. 2014. Structural diversity and function of

xyloglucan sidechain substituents. Plants 3: 526–542.Schultz CJ, Johnson KL, Currie G, Bacic A. 2000. The classical arabinogalactan

protein gene family of Arabidopsis. Plant Cell 12: 1751–1767.Sechet J, Htwe S, Urbanowicz B, Agyeman A, Feng W, Ishikawa T, Colomes

M, Kumar KS, Kawai-Yamada M, Dinneny JR et al. 2018. Suppression of

Arabidopsis GGLT1 affects growth by reducing the L-galactose content and

New Phytologist (2019) � 2019 The Authors

New Phytologist� 2019 New Phytologist Trustwww.newphytologist.com

Research

NewPhytologist16

borate cross-linking of rhamnogalacturonan-II. The Plant Journal 96: 1036–1050.

Seguela-Arnaud M, Smith C, Uribe MC, May S, Fischl H, McKenzie N, Bevan

MW. 2015. The Mediator complex subunits MED25/PFT1 and MED8 are

required for transcriptional responses to changes in cell wall arabinose

composition and glucose treatment in Arabidopsis thaliana. BMC Plant Biology15: 215.

Seifert GJ, Barber C, Wells B, Dolan L, Roberts K. 2002. Galactose

biosynthesis in Arabidopsis: genetic evidence for substrate channeling

from UDP-D-galactose into cell wall polymers. Current Biology 12:

1840–1845.Seifert GJ, Roberts K. 2007. The biology of arabinogalactan proteins. AnnualReview of Plant Biology 58: 137–161.

Sherson S, Gy I, Medd J, Schmidt R, Dean C, Kreis M, Lecharny A,

Cobbett C. 1999. The arabinose kinase, ARA1, gene of Arabidopsis is a

novel member of the galactose kinase gene family. Plant MolecularBiology 39: 1003–1012.

Shi H, Kim Y, Guo Y, Stevenson B, Zhu J-K. 2003. The Arabidopsis SOS5 locusencodes a putative cell surface adhesion protein and is required for normal cell

expansion. Plant Cell 15: 19–32.Showalter AM. 1993. Structure and function of plant cell wall proteins. PlantCell 5: 9–23.

Showalter AM, Basu D. 2016. Extensin and arabinogalactan-protein

biosynthesis: glycosyltransferases, research challenges, and biosensors. Frontiersin Plant Science 7: 814.

van Hengel AJ, Roberts K. 2002. Fucosylated arabinogalactan-proteins are

required for full root cell elongation in Arabidopsis. The Plant Journal 32: 105–113.

Velasquez SM, Ricardi MM, Dorosz JG, Fernandez PV, Nadra AD, Pol-Fachin

L, Egelund J, Gille S, Harholt J, Ciancia M et al. 2011. O-Glycosylated cellwall proteins are essential in root hair growth. Science 332: 1401–1403.

Worden N, Park E, Drakakaki G. 2012. Trans-Golgi network: an intersection of

trafficking cell wall components. Journal of Integrative Plant Biology 54: 875–886.

Supporting Information

Additional Supporting Information may be found online in theSupporting Information section at the end of the article.

Fig. S1 Identification of 20-2 mutant that is defective in rootelongation under salt stress.

Fig. S2 Mutations in MED25, MED8, and PRL1 genes slightlyincrease the root length of the mur4/hsr8 mutants under saltstress.

Fig. S3 The disrupted cell–cell adhesion in the roots of themur4-1 mutant is rescued by applying boric acid and gum arabic.

Fig. S4 mur4 mutant shows more burst cells than wild-typeunder high salinity.

Fig. S5 The isoforms of MUR4 do not form homodimers orheterodimers with each other.

Fig. S6 Salt stress-induced upregulation of the three paralogousgenes ofMUR4 depends on ABA-mediated signaling pathway.

Fig. S7 Disruption of each paralogous gene of MUR4 does notaffect root elongation under salt stress.

Table S1 Primers used in this study.

Table S2 IP-MS data generated from transgenic plants expressing35S::MUR4–GFP or 35S::GFP.

Please note: Wiley Blackwell are not responsible for the contentor functionality of any Supporting Information supplied by theauthors. Any queries (other than missing material) should bedirected to the New Phytologist Central Office.

New Phytologist is an electronic (online-only) journal owned by the New Phytologist Trust, a not-for-profit organization dedicatedto the promotion of plant science, facilitating projects from symposia to free access for our Tansley reviews and Tansley insights.

Regular papers, Letters, Research reviews, Rapid reports and both Modelling/Theory and Methods papers are encouraged. We are committed to rapid processing, from online submission through to publication ‘as ready’ via Early View – our average timeto decision is <26 days. There are no page or colour charges and a PDF version will be provided for each article.

The journal is available online at Wiley Online Library. Visit www.newphytologist.com to search the articles and register for tableof contents email alerts.

If you have any questions, do get in touch with Central Office (np-centraloffice@lancaster.ac.uk) or, if it is more convenient,our USA Office (np-usaoffice@lancaster.ac.uk)

For submission instructions, subscription and all the latest information visit www.newphytologist.com

� 2019 The Authors

New Phytologist� 2019 New Phytologist TrustNew Phytologist (2019)

www.newphytologist.com

NewPhytologist Research 17